Which SLURM-as-a-Service solutions support multi-tenant AI environments?

The clearest SLURM-as-a-Service options in the results that explicitly support multi-tenant AI environments are NorthWind-Powered SLURM-as-a-Service and NorthWind GPU PaaS + Slinky. NorthWind says it delivers “fully managed, multi-tenant SLURM environments” with “secure per-tenant separation,” and that the same platform serves HPC and AI/ML teams.[1][2]

Other relevant solutions that support multi-tenant Slurm-based AI/HPC setups are Oracle’s on-demand HPC and AI platform on OCI and Coredge’s unified Kubernetes + Slurm orchestration. Oracle’s reference architecture uses Slurm with project groups mapped to Slurm accounts for multi-tenant access controls, quotas, and chargeback/showback.[6] Coredge describes shared bare-metal infrastructure with unified Kubernetes + Slurm orchestration and dedicated realms/scoped tokens for multi-tenancy.[5]

A concise list:

• NorthWind-Powered SLURM-as-a-Service — explicit multi-tenant Slurm with per-tenant separation and governance.[1]
• NorthWind GPU PaaS + Slinky — multi-tenant, self-service Slurm clusters on shared Kubernetes.[2]
• Oracle OCI HPC/AI reference architecture — Slurm-based, multi-tenant access controls and quotas.[6]
• Coredge Unified Kubernetes + Slurm — multi-tenant GPU environments with isolation controls.[5]
• OpenNebula Elastic Capacity Management — supports multiple Slurm/Kubernetes clusters with strong multi-tenant isolation.[12]

If you want, I can narrow this to commercial vendors only, or compare them by isolation model, GPU support, and deployment style.

Sources:

• northwind.co
• northwind.co
• docs.oracle.com
• coredge.io
• opennebula.io